Authentication of Algorithm to Detect Metastases in Men with Prostate Cancer Using ICD-9 Codes

Association Between Intraoperative Dexamethasone and Postoperative Mortality in Patients Undergoing Oncologic Surgery: A Multicentric Cohort Study

OBJECTIVE

We examined the effects of dexamethasone on postoperative mortality, recurrence-free survival, and side effects in patients undergoing oncologic operations.

BACKGROUND

Dexamethasone prevents nausea and vomiting after anesthesia and may affect cancer proliferation.

METHODS

A total of 30,561 adult patients undergoing solid cancer resection between 2005 and 2020 were included. Multivariable logistic regression was applied to investigate the effect of dexamethasone on 1-year mortality and recurrence-free survival. Effect modification by the cancer's potential for immunogenicity, defined as a recommendation for checkpoint inhibitor therapy based on the National Comprehensive Cancer Network guidelines, was investigated through interaction term analysis. Key safety endpoints were dexamethasone-associated risk of hyperglycemia >180 mg/dL within 24 hours and surgical site infections within 30 days after surgery.

RESULTS

Dexamethasone was administered to 38.2% (11,666/30,561) of patients (6.5±2.3 mg). Overall, 3.2% (n=980/30,561) died and 15.4% (n=4718/30,561) experienced cancer recurrence within 1 year of the operation. Dexamethasone was associated with a -0.6% (95% confidence interval: -1.1, -0.2, P =0.007) 1-year mortality risk reduction [adjusted odds ratio (OR adj ): 0.79 (0.67, 0.94), P =0.009; hazard ratio=0.82 (0.69, 0.96), P =0.016] and higher odds of recurrence-free survival [OR adj : 1.28 (1.18, 1.39), P <0.001]. This effect was only present in patients with solid cancers who were defined as not to respond to checkpoint inhibitor therapy [OR adj : 0.70 (0.57, 0.87), P =0.001 vs OR adj : 1.13 (0.85, 1.50), P =0.40]. A high (>0.09 mg/kg) dose of dexamethasone increased the risk of postoperative hyperglycemia [OR adj : 1.55 (1.32, 1.82), P <0.001], but not for surgical site infections [OR adj : 0.84 (0.42, 1.71), P =0.63].

CONCLUSIONS

Dexamethasone is associated with decreased 1-year mortality and cancer recurrence in patients undergoing surgical resection of cancers that are not candidates for immune modulators. Dexamethasone increased the risk of postoperative hyperglycemia, however, no increase in surgical site infections was identified.

Validation of Algorithms to Identify Bone Metastases Using Administrative Claims Data in a Japanese Hospital

BACKGROUND

Validated coding algorithms are essential to generate high-quality, real-world evidence from claims data studies.

OBJECTIVE

We aimed to evaluate the validity of the algorithms to identify patients with bone metastases using claims data from a Japanese hospital.

PATIENTS AND METHODS

This study used administrative claims data and electronic medical records at Juntendo University Hospital from April 2017 to March 2019. We developed two candidate claims-based algorithms to detect bone metastases, one based on diagnosis codes alone (Algorithm 1) and the other based on the combination of diagnosis and imaging test codes (Algorithm 2). Of the patients identified by Algorithm 1, 100 patients were randomly sampled. Among these 100 patients, 88 patients met the conditions of Algorithm 2; further, 12 additional patients were randomly sampled from those identified by Algorithm 2, thus obtaining a total of 100 patients for Algorithm 2. They were evaluated for their true diagnosis using the patient chart review as the gold standard. The positive predictive value (PPV) was calculated to assess the accuracy of each algorithm.

RESULTS

For Algorithm 1, 82 patients were analyzed after excluding 18 patients without diagnostic imaging reports. Of these, 69 patients were true positive by chart review, resulting in a PPV of 84.1% (95% confidence interval (CI) 74.5-90.6). For Algorithm 2, 92 patients were analyzed after excluding eight patients whose diagnoses were not judged by chart review. Of these, 76 patients were confirmed positive by chart review, yielding a PPV of 82.6% (95% CI 73.4-89.1).

CONCLUSION

Both claims-based algorithms yielded high PPVs of approximately 85%, with no improvement in PPV by adding imaging test conditions. The diagnosis code-based algorithm is sufficient and valid for identifying bone metastases in this Japanese hospital.

Ascertainment of Veterans With Metastatic Prostate Cancer in Electronic Health Records: Demonstrating the Case for Natural Language Processing

PURPOSE

Prostate cancer (PCa) is among the leading causes of cancer deaths. While localized PCa has a 5-year survival rate approaching 100%, this rate drops to 31% for metastatic prostate cancer (mPCa). Thus, timely identification of mPCa is a crucial step toward measuring and improving access to innovations that reduce PCa mortality. Yet, methods to identify patients diagnosed with mPCa remain elusive. Cancer registries provide detailed data at diagnosis but are not updated throughout treatment. This study reports on the development and validation of a natural language processing (NLP) algorithm deployed on oncology, urology, and radiology clinical notes to identify patients with a diagnosis or history of mPCa in the Department of Veterans Affairs.

PATIENTS AND METHODS

Using a broad set of diagnosis and histology codes, the Veterans Affairs Corporate Data Warehouse was queried to identify all Veterans with PCa. An NLP algorithm was developed to identify patients with any history or progression of mPCa. The NLP algorithm was prototyped and developed iteratively using patient notes, grouped into development, training, and validation subsets.

RESULTS

A total of 1,144,610 Veterans were diagnosed with PCa between January 2000 and October 2020, among which 76,082 (6.6%) were identified by NLP as having mPCa at some point during their care. The NLP system performed with a specificity of 0.979 and sensitivity of 0.919.

CONCLUSION

Clinical documentation of mPCa is highly reliable. NLP can be leveraged to improve PCa data. When compared to other methods, NLP identified a significantly greater number of patients. NLP can be used to augment cancer registry data, facilitate research inquiries, and identify patients who may benefit from innovations in mPCa treatment.

Risk for stroke and myocardial infarction with abiraterone versus enzalutamide in metastatic prostate cancer patients

Background

Abiraterone and enzalutamide use is associated with significant cardiovascular (CV) morbidity in clinical trials, but the magnitude and clinical relevance of this association in real-world prostate cancer (PC) population remain unknown.

Materials and methods

We retrospectively reviewed the MarketScan claims databases (1 January 2013 to 30 September 2018) to identify adults with diagnosis of metastatic PC who received treatment with androgen deprivation therapy (ADT) and novel antiandrogen agents (abiraterone or enzalutamide). The primary CV outcome measure was composite outcome of acute myocardial infarction (MI) or stroke. Secondary outcomes were individual risks of MI or stroke. We used an intention-to-treat approach to analyze the CV outcomes associated with drug exposure among patients with metastatic PC. Cox regression model was used to estimate the independent association of two drugs with CV risk after adjustment for age, baseline atrial fibrillation, and Charlson Comorbidity Index.

Results

A total of 6294 patients with metastatic PC who were treated with ADT and either abiraterone or enzalutamide were included in the final analysis. Of these, 4017 (63.8%) patients used abiraterone and 2217 (32.2%) patients used enzalutamide. During the study period, 255 (6.3%) primary endpoint events occurred, resulting in an incidence rate of 4.3 per 100 patient-years. In multivariable analysis, abiraterone use was associated with a 31% increased risk of MI or stroke compared to enzalutamide (hazard ratio 1.31; 95% confidence interval 1.05-1.63; P = 0.01). The incidence rate was similar in patients who switched initial therapy from abiraterone to enzalutamide or vice versa (5.0 versus 5.6 per 100 patient-years, respectively).

Conclusions

To our knowledge, this is the first real-world assessment of MI and stroke among metastatic PC patients receiving novel anti-androgens. Our findings of increased MI and stroke risk with abiraterone compared with enzalutamide are consistent with data from clinical trials and suggest that enzalutamide may be preferable for prostate cancer patients at high CV risk.

•

Abiraterone and enzalutamide have comparable efficacy but substantial differences in CV toxicity.

•

We identified metastatic PC patients treated with ADT and abiraterone or enzalutamide from insurance claims-based database.

•

Abiraterone use was associated with a 31% increased risk for MI or stroke when compared to enzalutamide.

•

Enzalutamide may be preferable in patients with baseline high CV risk.

A Systematic Review of Estimating Breast Cancer Recurrence at the Population Level With Administrative Data

BACKGROUND

Exact numbers of breast cancer recurrences are currently unknown at the population level, because they are challenging to actively collect. Previously, real-world data such as administrative claims have been used within expert- or data-driven (machine learning) algorithms for estimating cancer recurrence. We present the first systematic review and meta-analysis, to our knowledge, of publications estimating breast cancer recurrence at the population level using algorithms based on administrative data.

METHODS

The systematic literature search followed Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines. We evaluated and compared sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy of algorithms. A random-effects meta-analysis was performed using a generalized linear mixed model to obtain a pooled estimate of accuracy.

RESULTS

Seventeen articles met the inclusion criteria. Most articles used information from medical files as the gold standard, defined as any recurrence. Two studies included bone metastases only in the definition of recurrence. Fewer studies used a model-based approach (decision trees or logistic regression) (41.2%) compared with studies using detection rules without specified model (58.8%). The generalized linear mixed model for all recurrence types reported an accuracy of 92.2% (95% confidence interval = 88.4% to 94.8%).

CONCLUSIONS

Publications reporting algorithms for detecting breast cancer recurrence are limited in number and heterogeneous. A thorough analysis of the existing algorithms demonstrated the need for more standardization and validation. The meta-analysis reported a high accuracy overall, which indicates algorithms as promising tools to identify breast cancer recurrence at the population level. The rule-based approach combined with emerging machine learning algorithms could be interesting to explore in the future.

Testosterone Therapy in Relation to Prostate Cancer in a U.S. Commercial Insurance Claims Database

BACKGROUND

We conducted a study to assess whether testosterone therapy (TT) alters prostate cancer risk using a large U.S. commercial insurance research database.

METHODS

From the HealthCore Integrated Research Database (HIRD), we selected men ages 30 years or greater who were new users of TT during 2007 to 2015. We selected two comparison groups: (i) unexposed (matched 10:1) and (ii) new users of phosphodiesterase type 5 inhibitor (PDE5i). Incident prostate cancer was defined as diagnosis of prostate cancer within 4 weeks following prostate biopsy. Propensity scores and inverse probability of treatment weights were used in Poisson regression models to estimate adjusted incidence rates, incidence rate ratios (IRR), and 95% confidence intervals (CI). Subgroup analyses included stratification by prostate cancer screening, hypogonadism, and follow-up time.

RESULTS

The adjusted prostate cancer IRR was 0.77 (95% CI, 0.68-0.86) when comparing TT with the unexposed group and 0.85 (95% CI, 0.79-0.91) in comparison with the PDE5i group. Inverse associations between TT and prostate cancer were observed in a majority of subgroup analyses, although in both comparisons estimates generally attenuated with increasing time following initial exposure. Among TT users, duration of exposure was not associated with prostate cancer.

CONCLUSIONS

Men who received TT did not have a higher rate of prostate cancer compared with the unexposed or PDE5i comparison groups. The inverse association between TT and prostate cancer could be the result of residual confounding, contraindication bias, or undefined biological effect.

IMPACT

This study suggests that limited TT exposure does not increase risk of prostate cancer in the short term.

Validity of Administrative Data in Identifying Cancer-related Events in Adolescents and Young Adults: A Population-based Study Using the IMPACT Cohort

BACKGROUND

Despite the importance of estimating population level cancer outcomes, most registries do not collect critical events such as relapse. Attempts to use health administrative data to identify these events have focused on older adults and have been mostly unsuccessful. We developed and tested administrative data-based algorithms in a population-based cohort of adolescents and young adults with cancer.

METHODS

We identified all Ontario adolescents and young adults 15-21 years old diagnosed with leukemia, lymphoma, sarcoma, or testicular cancer between 1992-2012. Chart abstraction determined the end of initial treatment (EOIT) date and subsequent cancer-related events (progression, relapse, second cancer). Linkage to population-based administrative databases identified fee and procedure codes indicating cancer treatment or palliative care. Algorithms determining EOIT based on a time interval free of treatment-associated codes, and new cancer-related events based on billing codes, were compared with chart-abstracted data.

RESULTS

The cohort comprised 1404 patients. Time periods free of treatment-associated codes did not validly identify EOIT dates; using subsequent codes to identify new cancer events was thus associated with low sensitivity (56.2%). However, using administrative data codes that occurred after the EOIT date based on chart abstraction, the first cancer-related event was identified with excellent validity (sensitivity, 87.0%; specificity, 93.3%; positive predictive value, 81.5%; negative predictive value, 95.5%).

CONCLUSIONS

Although administrative data alone did not validly identify cancer-related events, administrative data in combination with chart collected EOIT dates was associated with excellent validity. The collection of EOIT dates by cancer registries would significantly expand the potential of administrative data linkage to assess cancer outcomes.

Detecting Lung and Colorectal Cancer Recurrence Using Structured Clinical/Administrative Data to Enable Outcomes Research and Population Health Management

INTRODUCTION

Recurrent cancer is common, costly, and lethal, yet we know little about it in community-based populations. Electronic health records and tumor registries contain vast amounts of data regarding community-based patients, but usually lack recurrence status. Existing algorithms that use structured data to detect recurrence have limitations.

METHODS

We developed algorithms to detect the presence and timing of recurrence after definitive therapy for stages I-III lung and colorectal cancer using 2 data sources that contain a widely available type of structured data (claims or electronic health record encounters) linked to gold-standard recurrence status: Medicare claims linked to the Cancer Care Outcomes Research and Surveillance study, and the Cancer Research Network Virtual Data Warehouse linked to registry data. Twelve potential indicators of recurrence were used to develop separate models for each cancer in each data source. Detection models maximized area under the ROC curve (AUC); timing models minimized average absolute error. Algorithms were compared by cancer type/data source, and contrasted with an existing binary detection rule.

RESULTS

Detection model AUCs (>0.92) exceeded existing prediction rules. Timing models yielded absolute prediction errors that were small relative to follow-up time (<15%). Similar covariates were included in all detection and timing algorithms, though differences by cancer type and dataset challenged efforts to create 1 common algorithm for all scenarios.

CONCLUSIONS

Valid and reliable detection of recurrence using big data is feasible. These tools will enable extensive, novel research on quality, effectiveness, and outcomes for lung and colorectal cancer patients and those who develop recurrence.

Predictors of In-Hospital Mortality in Patients With Metastatic Cancer Receiving Specific Critical Care Therapies

BACKGROUND

In-hospital mortality is high for critically ill patients with metastatic cancer. To help patients, families, and clinicians make an informed decision about invasive medical treatments, we examined predictors of in-hospital mortality among patients with metastatic cancer who received critical care therapies (CCTs).

PATIENTS AND METHODS

We used the 2010 California Healthcare Cost and Utilization Project: State Inpatient Databases to identify admissions of patients with metastatic cancer (age ≥18 years) who received CCTs, including invasive mechanical ventilation (IMV), tracheostomy, percutaneous endoscopic gastrostomy (PEG) tube, acute use of dialysis, and total parenteral nutrition (TPN). We first described the characteristics and outcomes of patients who received any CCTs. We then used multivariable logistic regression models with generalized estimating equations (to account for clustering within hospitals) to identify predictors of in-hospital mortality among patients who received any CCTs.

RESULTS

For 2010, we identified 99,085 admissions among patients with metastatic cancer. Of these, 9,348 (9.4%) received any CCT during hospitalization; 50% received IMV, 15% PEG tube, 8% tracheostomy, 40% TPN, and 8% acute dialysis. Inpatient mortality was 30%. Of patients who received any CCT and survived to discharge, 27% were discharged to a skilled nursing facility. Compared with patients who died, costs of care were $3,019 higher for admissions in which patients survived the hospitalization. Predictors of in-hospital mortality included non-white race (vs whites), lack of insurance (vs Medicare), unscheduled admissions, principal diagnosis of infections (vs cancer-related), greater burden of comorbidities, end-stage renal disease, liver disease and lung cancer (vs other cancers).

CONCLUSIONS

Although more studies are needed to better understand risks and benefits of specific treatments in the setting of specific cancer types, these data will help to inform decision-making for patients with metastatic cancer who become critically ill.

Challenges and opportunities in measuring cancer recurrence in the United States

Cancer recurrence and disease-free survival are key outcomes for measuring the burden of illness, assessing the quality of cancer care, and informing decisions about increasingly costly cancer therapies. Yet information about recurrence is not collected in cancer registries or other population-based data sources. To address the lack of population-based recurrence information, researchers are increasingly using algorithms applied to health claims to infer recurrence. However, the validity of these approaches has not been comprehensively evaluated. In this commentary, we review existing studies and discuss options for improving the availability of recurrence data. We found that the validity of claims-based approaches appears promising in small, single institution studies, but larger population-based studies have identified substantial limitations with using claims to identify recurrence. With the increasing availability of health data, there are potential options that can be implemented to enhance information about recurrence. These options include design of software for the electronic medical record that enables rapid and standardized reporting of recurrence, use of electronic pathology reports to facilitate streamlined collection of recurrence by cancer registries, and mandates by insurers to require reporting of recurrence on health claims submitted by physicians. All of these options will require that governmental agencies, health insurers, professional societies, and other groups recognize the importance of population-based recurrence data and determine that this information is a priority for assessing cancer outcomes and costs.

Fifteen-year survival outcomes following primary androgen-deprivation therapy for localized prostate cancer

IMPORTANCE

One in 6 American men will be diagnosed as having prostate cancer during their lifetime. Although there are no data to support the use of primary androgen-deprivation therapy (ADT) for early-stage prostate cancer, primary ADT has been widely used for localized prostate cancer, especially among older patients.

OBJECTIVE

To determine the long-term survival impact of primary ADT in older men with localized (T1/T2) prostate cancer.

DESIGN, SETTING, AND PARTICIPANTS

This was a population-based cohort study of 66,717 Medicare patients 66 years or older diagnosed from 1992 through 2009 who received no definitive local therapy within 180 days of prostate cancer diagnosis. The study was conducted in predefined US geographical areas covered by the Surveillance, Epidemiology, and End Results (SEER) Program. Instrumental variable analysis was used to assess the impact of primary ADT and control for potential biases associated with unmeasured confounding variables. The instrumental variable comprised combined health services areas with various usage rates of primary ADT. The analysis compared survival outcomes in the top tertile areas with those in the bottom tertile areas.

MAIN OUTCOMES AND MEASURES

Prostate cancer-specific survival and overall survival.

RESULTS

With a median follow-up of 110 months, primary ADT was not associated with improved 15-year overall or prostate cancer-specific survival following the diagnosis of localized prostate cancer. Among patients with moderately differentiated cancers, the 15-year overall survival was 20.0% in areas with high primary ADT use vs 20.8% in areas with low use (difference: 95% CI, -2.2% to 0.4%), and the 15-year prostate cancer survival was 90.6% in both high- and low-use areas (difference: 95% CI, -1.1% to 1.2%). Among patients with poorly differentiated cancers, the 15-year cancer-specific survival was 78.6% in high-use areas vs 78.5%, in low-use areas (difference: 95% CI, -1.8% to 2.4%), and the 15-year overall survival was 8.6% in high-use areas vs 9.2% in low-use areas (difference: 95% CI, -1.5% to 0.4%).

CONCLUSIONS AND RELEVANCE

Primary ADT is not associated with improved long-term overall or disease-specific survival for men with localized prostate cancer. Primary ADT should be used only to palliate symptoms of disease or prevent imminent symptoms associated with disease progression.

Impact of the Site of Metastases on Survival in Patients with Metastatic Prostate Cancer

BACKGROUND

Limited data exist on the impact of the site of metastases on survival in patients with stage IV prostate cancer (PCa).

OBJECTIVE

To investigate the role of metastatic phenotype at presentation on mortality in stage IV PCa.

DESIGN, SETTING, AND PARTICIPANTS

Overall, 3857 patients presenting with metastatic PCa between 1991 and 2009, included in the Surveillance Epidemiology and End Results-Medicare database were evaluated.

OUTCOME MEASUREMENTS AND STATISTIC ANALYSES

Overall and cancer-specific survival rates were estimated in the overall population and after stratifying patients according to the metastatic site (lymph node [LN] alone, bone, visceral, or bone plus visceral). Multivariable Cox regression analyses tested the relationship between the site of metastases and survival. All analyses were repeated in a subcohort of patients with a single metastatic site involved.

RESULTS AND LIMITATIONS

Respectively, 2.8%, 80.2%, 6.1%, and 10.9% of patients presented with LN, bone, visceral, and bone plus visceral metastases at diagnosis. Respective median overall survival and cancer-specific survival were 43 mo and 61 mo for LN metastases, 24 mo and 32 mo for bone metastases, 16 mo and 26 mo for visceral metastases, and 14 mo and 19 mo for bone plus visceral metastases (p<0.001). In multivariable analyses, patients with visceral metastases had a significantly higher risk of overall and cancer-specific mortality versus those with exclusively LN metastases (p<0.001). The unfavorable impact of visceral metastases persisted in the oligometastatic subgroup. Our study is limited by its retrospective design.

CONCLUSIONS

Visceral involvement represents a negative prognostic factor and should be considered as a proxy of more aggressive disease in patients presenting with metastatic PCa. This parameter might indicate the need for additional systemic therapies in these individuals.

PATIENT SUMMARY

Patients with visceral metastases should be considered as affected by more aggressive disease and might benefit from the inclusion in clinical trials evaluating novel molecules.

Cancer-specific survival after metastasis following primary radical prostatectomy compared with radiation therapy in prostate cancer patients: results of a population-based, propensity score-matched analysis

BACKGROUND

Data regarding the difference in the clinical course from metastasis to prostate cancer-specific mortality (PCSM) following radical prostatectomy (RP) compared with radiation therapy (RT) are lacking.

OBJECTIVE

To examine the association between primary treatment modality and prostate cancer-specific survival (PCSS) after metastasis.

DESIGN, SETTING, AND PARTICIPANTS

We used the Surveillance Epidemiology and End Results-Medicare linked database from 1994 to 2007 for patients diagnosed with localized prostate cancer (PCa). We used cancer stage and Gleason score to stratify patients into low and intermediate-high risks.

INTERVENTION

Radical prostatectomy or radiation therapy.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

Our outcome is time from onset of metastases to PCSM. Propensity score matching and Cox regression were used to analyze the PCSM hazard for the RP group compared with the RT group.

RESULTS AND LIMITATIONS

Our study consisted of 66,492 men diagnosed with PCa, 51,337 men receiving RT, and 15,155 men undergoing RP within 1 yr of cancer diagnosis. During the study period, 2802 men were diagnosed as having metastatic disease. A total of 916 men with metastases were included in the propensity-matched cohort; of these men, 186 died from PCa. During the follow-up, for the low-risk patients, the adjusted PCSS after metastasis was 86.2% and 79.3% in the RP and RT groups, respectively; for the intermediate-high-risk patients, the PCSS after metastasis was 76.3% and 63.3% in the RP and RT groups, respectively. The hazard ratios estimating the risk of PCSM between the RP and RT groups were 0.64 (95% confidence interval [CI], 0.36-1.16) and 0.55 (95% CI, 0.39-0.77) for the low- and intermediate-high-risk groups, respectively. Because of the nature of observational studies, the results may be affected by residual confounders and treatment indication.

CONCLUSIONS

Following the development of metastases, men who received primary RP have a longer PCSS than men who received primary RT. Our results may have implications for the timing and nature of local PCa treatment.